Article Cite This: Acc. Chem. Res. 2019, 52, 170−179
pubs.acs.org/accounts
Iodide Photoredox and Bond Formation Chemistry Ludovic Troian-Gautier, Wesley B. Swords, and Gerald J. Meyer*
Acc. Chem. Res. 2019.52:170-179. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/16/19. For personal use only.
Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, United States
CONSPECTUS: Iodide redox chemistry is intimately coupled with the formation and breaking of chemical bonds that are relevant to emerging solar energy technologies. In this Account, recent advances in dye-sensitized iodide oxidation chemistry in organic solutions are described. Here RuII sensitizers with high cationic charge, tuned reduction potentials, and specific iodide receptor site(s) are shown to self-assemble in organic solvents and yield structures that rapidly oxidize iodide and generate I−I bonds when illuminated with visible light. These studies provided new insights into the fascinating behavior of our most polarizable and easily oxidized monatomic anion. Sensitized iodide photo-oxidation in CH3CN solutions consists of two mechanistic steps. In the first step, an excited-state sensitizer oxidizes iodide (I−) to an iodine atom (I•) through diffusional encounters. The second step involves the reaction of I• with I− to form the I−I bond of diiodide, I2•−. The overall reaction converts a green photon into about 1.64 eV of free energy in the form of I2•− and the reduced sensitizer. The free energy is only transiently available, as back-electron transfer to yield ground-state products is quantitative. Interestingly, when the free energy change is near zero, iodide photo-oxidation occurs rapidly with rate constants near the diffusion limit, i.e., >1010 M−1 s−1. Such rapid reactivity is in line with anecdotal knowledge that iodide is an outstanding electron donor and is indicative of adiabatic electron transfer through an inner-sphere mechanism. In low-dielectric-constant solvents, dicationic RuII sensitizers were found to form tight ion pairs with iodide. Diimine ligands with additional cationic charge, or “binding pockets” that recognize halides, have been utilized to position one or more halides at specific locations about the sensitizer before light absorption. Diverse photochemical reactions observed with these supramolecular assemblies range from the photorelease of halides to the formation of I−I bonds where both iodides present in the ground-state assembly react. Natural population analysis through density functional theory calculations accurately predicts the site(s) of iodide ion-pairing and provides information on the associated free energy change. The ability to direct light-driven bond formation in these ionic assemblies is extended to chloride and bromide ions. The structure−property relationships identified, and those that continue to emerge, may one day allow for the rational design of molecules and materials that drive desired halide transformations when illuminated with light. dye-sensitized iodide oxidation with RuII bipyridine sensitizers in organic solutions. An early discovery was that the yields and kinetics for iodide photo-oxidation were much more favorable in organic solvents than they were in water.9 In polar solvents like acetonitrile, bimolecular reaction chemistry was operative that proceeded through diffusional encounters between the excited state and iodide. Kinetic and thermodynamic data strongly suggested that iodide forms adducts within the excited-state encounter complex that result in adiabatic electron transfer and an innersphere mechanism. Efforts to characterize a ground-state “encounter complex” through ion-pairing with cationic sensitizers in low-dielectric-constant solvents were successful
1. INTRODUCTION Iodide has recently gained considerable attention for solar energy conversion applications that include perovskite photovoltaics,1 dye-sensitized solar cells,2,3 quantum dot passivated cells,4 and H2 gas generation through HI splitting.5 As an outer-sphere electron donor, iodide has historically been the preferred anion that gives rise to distinct charge transfer absorption bands that report on solvent polarity.6 Iodide is a soft Lewis basic anion with a diffuse electron cloud that promotes strong electronic coupling at internuclear distances greater than 4 Å.7 The oxidation of iodide to an iodine atom is intimately coupled with the subsequent I−I bond formation and disproportionation chemistry. The polyiodide intermediates and products gain Lewis acidic character that gives rise to a very rich chemistry that can be initiated by the one-electron photoredox chemistry of iodide.7,8 This Account focuses on © 2018 American Chemical Society
Received: July 30, 2018 Published: December 20, 2018 170
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research and have enabled more detailed characterization of the sensitizer sites where iodide and other halides prefer to ionpair. An advantage of iodide sensitizer ion pairs for photoredox chemistry is that the ionic structure can be preassembled in the dark for desired light-initiated electron transfer and I−I bond formation while avoiding unwanted redox chemistry. In particular, the presence of H-bonding functional groups provides a method to coordinate and direct diffusing iodide ions to specific locations. Likewise, the use of sterically bulky substituents has also been utilized to prevent interactions with acidic hydrogen atoms on parts of the sensitizer where iodide coordination is unwanted. A question that naturally arises in the photochemistry of such supramolecular assemblies is which iodides are most easily oxidized: the ion-paired iodides, the Hbonded iodides, or the ones solvated in solution? Natural population analysis through density functional theory provides some direct information on this question that has been tested through experiment. Overall, supramolecular assembly through halide recognition and ion-pairing removes the diffusional limitations associated with bimolecular redox chemistry and allows for the rational design of sensitizers for desired applications.
Figure 1. (a) Structure of [Ru(bpz)2(deeb)]2+. (b) Time-resolved photoluminescence decays for [Ru(bpz)2(deeb)]2+* in the presence of iodide. The inset shows the corresponding Stern−Volmer plot for these data. (c) Absorption spectra of [Ru(bpz)2(deeb)]2+ (black) and its reduced form [Ru(bpz)2(deeb)]+ (red) and the photoluminescence spectrum of [Ru(bpz)2(deeb)]2+ (green). (d) Absorption difference spectra measured at the indicated delay times after pulsed 532 nm light excitation of a CH3CN solution containing [Ru(bpz)2(deeb)]2+ and 500 mM iodide. The inset shows a plot of the first-order rate constants for the formation of I2•− (black squares), the formation of [Ru(bpz)2(deeb)]+ (blue triangles), and excited-state decay (red circles) as functions of iodide concentration. Adapted from ref 13. Copyright 2009 American Chemical Society.
2. CLASSICAL RU DIIMINE COMPLEXES AS SENSITIZERS The metal-to-ligand charge transfer (MLCT) excited states of classical RuII diimine complexes have been widely utilized to sensitize iodide oxidation to visible light in organic solvents.10−20 These RuII sensitizers are ideal for fundamental studies because of their long-lived photoluminescent excited states, readily quantified and tuned formal reduction potentials, and high stability in adjacent oxidation states.
separate unsensitized ultraviolet light experiments that yielded I• in iodide solutions, which was found to react with the same second-order rate constant, kI = (2.4 ± 0.2) × 1010 M−1 s−1. The data were fully consistent with the mechanism shown in Figure 2. Upon light absorption, the diffusive MLCT excited state formed an encounter complex within which I− transferred an electron to the photoexcited complex to yield an iodine atom and the reduced ruthenium complex. About 4.2% of the iodine atoms did not undergo back-electron transfer within the encounter complex and instead escaped the solvent cage and reacted with iodide to yield I2•−. The reduced ruthenium complex and I2•− subsequently underwent back-electron transfer to yield ground-state products (kbet = 2.1 × 1010 M−1 s−1) in a fully reversible manner. Back-electron transfer occurred in kinetic competition with I2•− disproportionation (2I2•− → I3− + I−), which was not observed under these conditions but represents the predominant reaction pathway for this anion in dye-sensitized solar cells.2 In this photoredox chemistry, a green photon (2.3 eV) was converted to ∼1.64 eV of free energy in the transiently formed {[Ru(bpz−)(bpz)(deeb)]+, I2•−} redox equivalents. The observed rate constant, kobs, for electron transfer from iodide to [Ru(bpz)2(deeb)]2+* was close to the value expected for a diffusion-limited reaction. Equation 2 relates kobs to the rate constants for diffusion, kdiff, and electron transfer, ket, and the equilibrium constant for encounter complex formation, KA: 1 1 1 = + kobs kdiff KAket (2)
2.1. Diffusional Iodide Oxidation
Iodide photo-oxidation by [Ru(bpz)2(deeb)]2+* (bpz = 2,2′bipyrazine; deeb = 4,4′-(CO2CH2CH3)2-2,2′-bipyridine) was investigated in CH3CN at room temperature (Figure 1a).13,14 This complex was chosen as a potent photo-oxidant (E°(Ru2+*/+) = 1.36 V vs SCE), which was thought to be necessary on the basis of previous research in aqueous solution. The room-temperature photoluminescence intensity (PLI) of [Ru(bpz)2(deeb)]2+* and the excited-state lifetime (τ) were dramatically quenched by iodide. A traditional Stern−Volmer analysis of the PLI and τ yielded the same quenching constant, kq = KSV/τ0 = 6.5 × 1010 M−1 s−1, indicative of a dynamic process wherein electron transfer was preceded by a diffusional encounter of the excited state with iodide (Figure 1b and eq 1). PLI 0 τ = 0 = 1 + KSV[I−] = 1 + kq τ0[I−] PLI τ
(1)
The products of excited-state quenching were identified by the appearance of the characteristic absorption spectra of the reduced ruthenium complex, [Ru(bpz−)(bpz)(deeb)]+, and diiodide, I2•−, in the expected 1:1 ratio (Figure 1d). The reduced ruthenium complex appeared with the same rate constant as the excited-state decay, indicating that [Ru(bpz−)(bpz)(deeb)]+ was a primary photochemical product. In contrast, the appearance of I2•− was delayed. Iodine atom, I•, was proposed to be the primary photoproduct, which subsequently reacted with iodide to generate I2•−. This dyesensitized mechanism for I2•− formation was supported by
The kdiff value is easily calculated with standard equations and typically falls in the 109−1011 M−1 s−1 range in organic 171
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
Figure 2. (a) Mechanism for diffusional excited-state iodide oxidation and charge recombination. (b) Jablonski-type diagram of iodide oxidation by [Ru(bpz)2(deeb)]2+. (c) Uncorrected (kobs) and diffusion-corrected (kact) rate constants for the excited-state quenching of ruthenium polypyridyl complexes by iodide. Overlaid are fits to the Marcus theory.
Figure 3. (a) Structure of [Ru(deeb)3]2+. (b, c) Excited-state quenching of [Ru(deeb)3]2+ by iodide in (b) acetonitrile and (c) dichloromethane. The insets show the corresponding Stern−Volmer analyses.
solvents; a kdiff value of 6.4 × 1010 M−1 s−1 was calculated for this reactivity, which is similar to that measured experimentally. Within the encounter complex, electron transfer is a first-order reaction, so ket has units of s−1, while KA has units of M−1. The product of ket and KA, termed the activated rate constant, kact, with units of M−1 s−1, is often reported, as it corrects for diffusion (when necessary) and requires no assumptions about KA. The diffusion-limited reactivity of [Ru(bpz)2(deeb)]2+* with iodide indicated that weaker photo-oxidants would also be effective. Indeed, a family of eight different RuII sensitizers were also shown to generate iodine atoms and I2•− by the same
mechanism.12 The kact and kobs values are plotted against the free energy change for electron transfer, ΔG° (Figure 2c). The overlaid fit to Marcus theory provided a reorganization energy λ < 0.4 eV, which is an unrealistically small value; previous outer-sphere charge transfer studies have indicated that the solvent contribution alone should be >1 eV. Interestingly, even when ΔG° was near zero, the activated rate constants were large (∼1 × 1010 M−1 s−1). Such rapid reactivity and unrealistic parameters extracted from the fits to Marcus theory suggested that iodide formed an inner-sphere adduct, perhaps with the diimine ligand or directly to the Ru metal center, which provided an adiabatic electron transfer pathway in the 172
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
Figure 4. (a, b) Photoluminescence spectra of [Ru(dtb)2(dea)]2+* in dichloromethane with the addition of (a) up to 1 equiv of iodide and (b) more than 1 equiv of iodide. (c) Ground-state ion pair between [Ru(dtb)2(dea)]2+ and Cl−, Br− or I−. (d) Proposed mechanisms for the excitedstate quenching of [Ru(dtb)2(dea)]2+* by iodide. Electron localization in the excited state is emphasized by the blue shading of the dea ligand. Purple spheres represent iodine species. Adapted from ref 19. Copyright 2016 American Chemical Society.
quantified by 1H NMR, UV−vis, and PL spectroscopies in fluid solution and by X-ray crystallography in the solid state.22 Study of the homoleptic compounds was frustrated by the well-known ligand loss photochemistry that plagues this class of sensitizers,23 which was inhibited for the heteroleptic compounds, thereby enabling a more detailed study. A dramatic downfield shift of the 3- and 3′-H atoms of the bpy ligand was quantified by 1H NMR spectroscopy when Cl− was titrated into a CD2Cl2 solution of [Ru(bpy)2(deeb)]2+. These represent the most acidic H atoms of the bpy ligand. The resonances associated with the 6- and 6′-H atoms of the deeb ligand also shifted downfield, although to a lesser extent. When 0.1 equiv of Cl− was present, discrete resonances for the Cl− ion-paired and non-ion-paired species were not observed, indicative of rapid Cl− intermolecular ligand exchange on the NMR time scale. The downfield shift was proposed to result from attraction of H toward the Cl− that elongated the C−H bonds. This H-bonding to Cl− led to lowered electron density on the H atom and in turn deshielded 1H resonances. The strong preference for ion-pairing with the bpy ligand over the deeb ligand indicated that synthetic design could be used to control the site of ion-pairing.19,20,24
encounter complex. Strong coupling and adiabatic pathways were expected to decrease the activation barrier and the free energy change accompanying electron transfer.21 2.2. Static Iodide Oxidation
Insights into the chemical nature of the iodide−excited state encounter complex were garnered from ground-state studies in less polar solvents. For example, titration of iodide into a CH2Cl2 solution of [Ru(deeb)3]2+ led to significant changes in the MLCT absorption spectra indicative of ion-pairing that were not observed in the more polar acetonitrile.11 Iodide was also found to efficiently quench the [Ru(deeb)3]2+* excitedstate. Stern−Volmer analysis of the steady-state PLI showed upward curvature, which often indicates that both static and dynamic quenching mechanisms are operative. Indeed, timeresolved PL studies revealed a decreased initial amplitude with increased iodide concentration that provided an estimate of the ground-state equilibrium constant, Keq = 230 000 M−1. The excited-state was also dynamically quenched by iodide, with KSV = 41 000 M−1 (Figure 3). These constants were able to faithfully model the upward curvature of the steady-state PLI data with the fully integrated Stern−Volmer expression. In CH3CN, however, only dynamic quenching was observed, with KSV = 100 000 M−1. The static electron transfer was absent when an inert salt was present in CH2Cl2 at >100 mM, indicating that it was indeed due to the presence of ion pairs. To gain further insight into the location, equilibrium constant, and stoichiometry of ion-pairing, chloride was utilized as an “innocent” ion whose large reduction potential precluded excited-state electron transfer with these sensitizers. Chloride addition to dichloromethane solutions of the complexes [Ru(bpy)3−x(deeb)x]2+ (x = 0, 1, 2, 3) was
3. SENSITIZER SUPRAMOLECULAR ASSEMBLY The study of classical Ru diimine sensitizers revealed that iodide photo-oxidation was far more facile in organic solvents than in water, behavior attributed to inner-sphere adduct formation within the encounter complex.12 A natural extension was to utilize functional groups that specifically recognize and coordinate halide ions that had been identified previously in the anion sensing literature.25−27 Amide and alcohol functional 173
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
suggested that the ground- and excited-state equilibrium constants for halide association could differ.16,19 To better address the origin(s) of this behavior, two dicationic sensitizers were synthesized with the daea ligand, which coordinates halide ions in a manner similar to dea (Figure 5). The ancillary
groups as well as cationic tertiary amines have been shown to provide a means to position one or more iodide ions in specific locations about the sensitizer prior to light absorption. 3.1. Ligands that Promote H-Bonding with Iodide
The ligand 4,4′-diethanolamide-2,2′-bipyridine (dea) was utilized to recognize and bind halides through H-bonding interactions with the alcohol and amide groups.19 The complex [Ru(dtb)2(dea)]2+ was prepared, in which the tert-butyl groups were located at the 4- and 4′-positions of bipyridine (4,4′-ditert-butyl-2,2′-bipyridine) in an effort to inhibit iodide association with the 3- and 3′-H atoms (Figure 4). A 1:1 Ru:I stoichiometry was identified through 1H NMR titrations in CD2Cl2. These titrations indicated that the coordinated dea ligand recognizes halide ions in solution with the bipyridyl 3and 3′-H atoms and the N−H and O−H groups directing the interaction. Clear evidence for adduct formation was also present in the more polar CH3CN, although the equilibrium constants (Keq ∼ 104 M−1) were at least 2 orders of magnitude smaller than those measured in CH2Cl2 (Keq > 106 M−1), indicating the importance of electrostatic attraction.16 The addition of up to 1 equiv of iodide led to a surprising increase in the PL intensity and a blue-shifted PL spectrum for the 1:1 ion pair, {[Ru(dtb)2(dea)], X−}+* (Figure 4a). Chloride induced a more pronounced PLI enhancement, with nearly a 2-fold increase in the PL quantum yield that was traced to a more favorable nonradiative rate constant. Although speculative, the halide−hydrogen bonds were proposed to force a more planar orientation of the two pyridine rings of the dea ligand, resulting in a more favorable non-radiative rate constant. When more than 1 equiv of iodide was present, a dramatic quenching of the {[Ru(dtb)2(dea)], I−}+* excited state was evident (Figure 4b). Spectroscopic measurements revealed a purely dynamic electron transfer reaction. The three proposed quenching mechanisms shown in Figure 4d were considered. The first involved diffusional encounters of {[Ru(dtb)2(dea)], I−}+* with iodide to from an encounter complex in which excited-state electron transfer and I−I bond formation occurred in one concerted step. Such a “concerted pathway” avoids the high-energy I• intermediate and hence represents a more efficient means to produce I2•−. However, transient absorption data consistently showed that I2•− was a secondary photoproduct with strong evidence for an I• intermediate. Furthermore, when 1 equiv of chloride was first added as an “innocent” anion, the subsequent addition of excess iodide provided no evidence of I−Cl•− formation. Altogether, the data indicated that the iodide (or chloride) present within the dea ligand did not participate in excited-state quenching. Two of the proposed mechanisms involved dynamic quenching by either the free (pathway (ii)) or iodideassociated (pathway (iii)) excited state. A Debye−Hückel analysis of the quenching as a function of ionic strength revealed that the reaction took place with a monocationic excited state and a monatomic anion, revealing that pathway (iii) was indeed the operative one.19 Iodide association with the dea ligand was kinetically fast, and the H-bonded iodide was inert toward excited-state oxidation for reasons described below in section 3.4.
Figure 5. (a) Square scheme for ground- and excited-state equilibria of [Ru(dtb)2(daea)]2+ with chloride. The blue shading indicates the ligand on which the excited-state was localized in the MLCT excited state, and the green sphere represents chloride. (b) Transient photoluminescence spectra obtained 45 ns (purple squares) and at longer (purple to red) time delays after pulsed 500 nm laser excitation of [Ru(dtb)2(daea)]2+ (abbreviated Ru-daea) in the presence of 1 equiv of chloride. The photoluminescence spectrum in the absence of chloride is also given for reference (black triangles, dashed line). The bold black arrow indicates the spectral shift expected for chloride association. The colored arrow indicates the time-dependent spectral shift measured. Adapted from ref 28. Copyright 2018 American Chemical Society.
ligands provided control of the excited-state dipole orientation. In the complex [Ru(dtb)2(daea)]2+, the excited-state dipole was localized on the daea ligand, i.e., [RuIII(dtb)2(daea−)]2+* and hence toward the associated halide. In contrast, in [Ru(btfmb)2(daea)]2+ (btfmb = 4,4′-bis(trifluoromethyl)-2,2′bipyridine), the excited state was localized on the btfmb ligand, i.e., [RuIII(btfmb)(btfmb−)(daea)]2+* and hence away from the associated halide. The dipole orientation had little influence on the ground-state equilibrium constant (Keq ∼ 4 × 106 M−1) but induced a profound change in the excited-state equilibrium. Light excitation of {[Ru(dtb)2(daea)], Cl−}+ led to the time-dependent photoluminescence spectra shown in Figure 5 and the appearance of biexponential kinetics. These spectral data were consistent with the photorelease of Cl−. In a kinetic analysis similar to that reported for photoacids and
3.2. Ground- and Excited-State Equilibrium Constants Determine Halide Photorelease
Supramolecular assembly was recently shown to provide excited states that release chloride ions.28 Early studies 174
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research Scheme 1. Proposed Mechanism for Iodide Photo-oxidation by [Ru(deeb)2(tmam)]4+
Figure 6. (a) Proposed mechanism for visible-light excitation of the ter-ionic {Ru4+, 2I−} complex to yield an I−I bond. (b) Stern−Volmer analysis of the data with an overlaid fit. (c) 1H NMR spectra of [Ru(dtb)2(tmam)]4+ in deuterated acetone (black). The circles show the variations in the chemical shifts observed upon titration with TBAI between 0 equiv (bottom) and 5 equiv (top). Adapted from ref 20. Copyright 2018 American Chemical Society.
photobases,29 the excited-state equilibrium constant was shown to decrease by about a factor of 20 relative to the ground-state value. In contrast, light excitation of {[Ru(btfmb)2(daea)], Cl−}+ revealed no evidence of chloride photodissociation and instead displayed a 45-fold increase in the excited-state equilibrium constant. Hence, the photorelease of chloride (and presumably other halides) can be controlled at the molecular level through orientation of the excited-state dipole vector.
this end, a series of highly charged ruthenium polypyridyl complexes bearing the dicationic ligand 4,4′-bis(trimethylaminomethyl)-2,2′-bipyridine (tmam) bearing total charges of 4+ ([Ru(deeb) 2 (tmam)] 4 + ), 6+ ([Ru(tmam)2(deeb)]6+), and 8+ ([Ru(tmam)3]8+) were prepared.17 These compounds were found to ion-pair with iodide even in polar CH3CN solutions. The equilibrium constants increased with the cationic charge (Keq = 4000, 4400, and 7000 M−1, respectively). Iodide titrations with an 1H NMR assay revealed the presence of a “binding pocket” within the tmam ligand that included H-bonds with the 3- and 3′-H atoms of the tmam ligand and electrostatic interactions with the
3.3. Electrostatic Interactions for Light Driven I−I Bond Formation
Rapid electron transfer in ion pairs implicated Coulombic attraction as a key strategy for enhancing iodide oxidation. To 175
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
Figure 7. Contour plots of the calculated Coulombic work term, ωr (in eV), over the plane containing the tmam ligand in (a) the absence and (b) the presence of the iodide ion pair in (left) dichloromethane (ε = 8.9), (middle) acetone (ε = 20.7), and (right) acetonitrile (ε = 37.5). All atoms within 1 Å of this plane are shown as small colored dots. The tmam ligand is superimposed in white.
this field enables concerted iodide oxidation and bond formation. In summary, the rapid formation of an I−I bond was facilitated by the supramolecular assembly of two iodide ions and one photosensitizer as reactants prior to light absorption.
quaternary amines that provided resolution of the enantiotopic methylene −CH2− H atoms. The 1H NMR resonances associated with the deeb ligand were largely insensitive to iodide. When iodide was present in the binding pocket, rapid excited-state electron transfer occurred with a rate constant of >108 s−1 (Scheme 1). Translation of this ion-pair chemistry to the sensitized TiO2 interface was accomplished with [Ru(tmam)2(dcb)]6+ (dcb = 4,4′-(CO2H)2-bpy) and an anionic cobalt redox mediator.30 More recently, a “ter-ionic complex” comprising a tetracationic Ru−tmam complex with two iodide ions, {Ru4+, 2I−}, was found to produce I2•− upon light excitation in acetone (Figure 6).20 The 1H NMR spectra in Figure 6c revealed that one iodide was present in the tmam pocket and the second was closer to the Ru center near the 6,6′ and 5,5′ H atoms of the dtb and tmam ligands. The excited state was quenched by iodide, but remarkably, the quenching saturated at high iodide concentrations. This occurred even though continued diffusional quenching was expected with the ∼100 ns excited-state lifetime. This behavior is evident in the Stern− Volmer plots, which show upward curvature and saturation at high iodide concentrations (Figure 6b). A mechanism proposed to account for this behavior is shown in Figure 6a. The more strongly coupled iodide closest to the Ru center is oxidized first to generate an iodine atom. The neutral iodine atom subsequently forms a bond with the iodide in the tmam pocket in about 70 ns. The insensitivity of this value to the I− concentration indicated that this is an intraionic reaction that does not involve solvated iodide ions. Hence, this is not a concerted mechanism in which I−I bond formation and electron transfer occur in one step but rather represents a ground-state complex that forms a chemical bond between the two iodides when illuminated. The concerted mechanism likely is inhibited by the electron localized on the tmam ligand, which provides an electrostatic barrier for iodide diffusion. In future work, it will be of interest to localize the excited state on a remote ligand to test whether the absence of
3.4. Free Energy Change that Accompanies Excited-State Electron Transfer
The driving force for excited-state electron transfer, ΔG°et, has typically been determined with an expression similar to that of Rehm and Weller (eq 4):31 ΔGet° = [E°(I• / −) − E°(Ru n * /n − 1)]F + ΔGω
(4)
where F is Faraday’s constant. The first term in brackets is the difference in formal reduction potentials between iodine and the excited state, and the second term, ΔGω, is the free energy change associated with the work required to bring the reactants together and to separate the products, which is given by eq 5: ΔGω = ω(I•Ru n − 1) − ω(I−Ru n *) m
=
1 ∑ 1 [z(I•)zi(Run− 1) − z(I−)zi(Run*)] 4πε0εr i = 1 ai ,I− (5)
This equation provides a means for quantification through the charge (z) of the Ru and iodide species, the solvent dielectric constant (εr), and the electron-transfer distance (a). The expression shows that ΔGω is small in high-dielectric-constant solvents with weakly charged species and is often reasonably ignored in polar solvents like water and CH3CN. Work term calculations found in the literature often assume donor and acceptor point charges, an approximation that is most accurate when their separation is large.32 However, this approximation breaks down when the donor−acceptor distance is small and specific atomic interactions occur, such as when iodide is present in a ligand binding pocket. A more rigorous description of ΔGω involves pairwise summation of the atomic charge of each atom (zi) of the photosensitizer with 176
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
Another exciting research opportunity is to move up the halogen family to bromide and chloride.24,34−38 Periodic trends indicate that these ions will be more difficult to oxidize, yet the lessons learned from iodide teach that halide coordination chemistry and self-assembly can allow considerable tuning of the reaction driving force. Indeed, recent publications show that molecular excited states are able to oxidize Br− and Cl−.34−37 The photochemically produced chlorine or bromine atoms are potent oxidants with the energetics necessary to drive subsequent reactions such as water oxidation.39 A specific challenge encountered with this class of sensitizers was ligand loss photochemistry to form complexes of the type Ru(LL)(LL′)X2 (X = Br−, Cl−). In some cases, ligand loss was so extreme that the sample had to be constantly refreshed.34−36 Prior research has shown that this photochemistry occurs after activated crossing from the MLCT state to dissociative ligandfield excited states.23,40−45 Such photochemistry was absent for I− because electron transfer from iodide was rapid and outcompeted the activated crossing. The future identification of sensitizers that rapidly photo-oxidize chloride or bromide will also presumably circumvent this unwanted photochemical ligand loss and enable desired reaction chemistry with these more earth-abundant halides.
the individual atom−iodide distances (ai,I−). In initial attempts to accomplish this, natural population analysis with density functional theory was utilized to assign zi for each atom on the sensitizer.19 With these charges, the work term for the reactants, ω(I−Run*) (abbreviated ωr) was estimated for a range of iodide distances. The contour plots so produced for the first and second iodide ion pairs with [Ru(dtb)2(tmam)]4+ in DCM, acetone, and CH3CN in Figure 7 show the expected decrease in magnitude with solvent permittivity. Since ωr is estimated through a Coulomb’s law approximation for electrostatic association, the atoms with the more negative ωr values are the sites where iodide is predicted to associate. It was gratifying to find that these were indeed the sites identified by 1H NMR analysis. Since the charge of the iodine atom is zI• = 0, the work term contribution from the electron transfer product, ω(I•Run−1), to ΔGet° was expected to be zero. Hence, ΔG°et for iodide photo-oxidation was predicted to decrease by 500 meV when iodide sits in the tmam binding pocket. This was borne out experimentally in the ter-ionic complexes with [Ru(dtb)2(tmam)]4+*, where the iodide in the tmam pocket did not quench the excited state.19 The predicted location of the second iodide was also in good agreement with experiment. The associated work term was about half that of the first ion pair (ωr ≈ −250 meV). This too was in line with experiment and the expectation that the second iodide would be the first to be photo-oxidized.20 Hence, if multiple iodides were to associate with a cationic sensitizer in a titration experiment, the last one would be expected to be the most easily oxidized thermodynamically. This expectation with synthetic design provides a rational means to direct the order in which assembled ions are oxidized and covalently linked. Studies with a larger number of sensitizers in solvents that promote ion-pairing will allow these expectations and the utility of natural population analysis to be more rigorously tested.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Ludovic Troian-Gautier: 0000-0002-7690-1361 Gerald J. Meyer: 0000-0002-4227-6393 Notes
The authors declare no competing financial interest.
4. CONCLUSIONS AND OUTLOOK Iodide’s ability to rapidly transfer an electron to a molecular excited state under endergonic conditions provides a further indication of the strong electronic coupling provided by this large, polarizable anion. The use of low-dielectric-constant solvents and/or sensitizers with large cationic charge enabled the self-assembly of ground-state structures that photooxidized iodide and made I−I bonds within 70 ns. Alternatively, self-assembly directed halide ions to locations where light excitation resulted in their photorelease. The ability to rationally design such very different photochemical transformations is noteworthy. Furthermore, natural population analysis with density functional theory predicted the site(s) of iodide interactions that were confirmed by 1H NMR spectroscopy. An intriguing extension of the research described herein exists in the photochemistry of transition metal iodo complexes that can be viewed as inner-sphere iodide−metal adducts. For example, photon absorption by [BiI6]3− involved iodide-tobismuth ligand-to-metal charge transfer that formally yields an iodine atom and a reduced Bi center in the Franck−Condon excited state.33 Experimentally, light excitation of this complex with nanosecond time resolution resulted in iodide oxidation with an I2•−-ligated Bi adduct intermediate that subsequently recombined to yield ground-state [BiI6]3−. A wide range of metal iodo complexes exist in the literature, and their photochemical reactivity is ripe for further study.
Biographies Ludovic Troian-Gautier received his B.S. (2008), M.S (2010), and Ph.D. (2014) in chemistry from the Université Libre de Bruxelles in Belgium under the direction of Prof. C. Moucheron and Prof. A. Kirsch-De Mesmaeker. He undertook postdoctoral research at X4C on surface modification using calix[4]arene derivatives with Prof. I. Jabin and Dr. A. Mattiuzzi. He received the Belgian American Educational Foundation (BAEF) Fellowship and the Bourse d’Excellence WBI.World (2016−2018) to undertake postdoctoral research with Gerald J. Meyer at UNC. Wesley B. Swords completed a B.S. in biochemistry/chemistry from UC San Diego, where he prepared transition metal isocyanide clusters in the Joshua Figueroa group. He was an NSF Graduate Research Fellow in the Gerald Meyer group at UNC, where he earned his Ph.D in 2018. He is now a postdoctoral associate with Tehshik Yoon at UW-Madison. Gerald J. Meyer received his B.S. in chemistry and mathematics from SUNY-Albany and his Ph.D. from UW-Madison. After over 20 years at Johns Hopkins University, he is currently a Professor of Chemistry at UNC and Director of the UNC Energy Frontier Research Center entitled Alliance for Molecular Photoelectrode Design for Solar Fuels (AMPED). His interests include the excited states, photochemistry, and photoelectrochemistry of inorganic coordination compounds and solids. 177
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research
■
ACKNOWLEDGMENTS
■
REFERENCES
(18) Taheri, A.; Meyer, G. J. Temperature dependent iodide oxidation by MLCT excited states. Dalton Trans 2014, 43, 17856− 17863. (19) Troian-Gautier, L.; Beauvilliers, E. E.; Swords, W. B.; Meyer, G. J. Redox Active Ion-Paired Excited States Undergo Dynamic Electron Transfer. J. Am. Chem. Soc. 2016, 138, 16815−16826. (20) Wehlin, S. A. M.; Troian-Gautier, L.; Sampaio, R. N.; Marcélis, L.; Meyer, G. J. Ter-Ionic Complex that Forms a Bond Upon Visible Light Absorption. J. Am. Chem. Soc. 2018, 140, 7799−7802. (21) Sampaio, R. N.; Piechota, E. J.; Troian-Gautier, L.; Maurer, A. B.; Hu, K.; Schauer, P. A.; Blair, A. D.; Berlinguette, C. P.; Meyer, G. J. Kinetics teach that electronic coupling lowers the free-energy change that accompanies electron transfer. Proc. Natl. Acad. Sci. U. S. A. 2018, 115, 7248−7253. (22) Ward, W. M.; Farnum, B. H.; Siegler, M.; Meyer, G. J. Chloride Ion-Pairing with Ru(II) Polypyridyl Compounds in Dichloromethane. J. Phys. Chem. A 2013, 117, 8883−8894. (23) White, J. K.; Schmehl, R. H.; Turro, C. An overview of photosubstitution reactions of Ru(II) imine complexes and their application in photobiology and photodynamic therapy. Inorg. Chim. Acta 2017, 454, 7−20. (24) Troian-Gautier, L.; Wehlin, S. A. M.; Meyer, G. J. Photophysical Properties of Tetracationic Ruthenium Complexes and Their Ter-Ionic Assemblies with Chloride. Inorg. Chem. 2018, 57, 12232− 12244. (25) Liu, Y.; Sengupta, A.; Raghavachari, K.; Flood, A. H. Anion Binding in Solution: Beyond the Electrostatic Regime. Chem. 2017, 3, 411−427. (26) Evans, N. H.; Beer, P. D. Advances in Anion Supramolecular Chemistry: From Recognition to Chemical Applications. Angew. Chem., Int. Ed. 2014, 53, 11716−11754. (27) Beer, P. D.; Gale, P. A. Anion Recognition and Sensing: The State of the Art and Future Perspectives. Angew. Chem., Int. Ed. 2001, 40, 486−516. (28) Turlington, M. D.; Troian-Gautier, L.; Sampaio, R. N.; Beauvilliers, E. E.; Meyer, G. J. Ligand Control of Supramolecular Chloride Photorelease. Inorg. Chem. 2018, 57, 5624−5631. (29) O’Donnell, R. M.; Sampaio, R. N.; Li, G.; Johansson, P. G.; Ward, C. L.; Meyer, G. J. Photoacidic and Photobasic Behavior of Transition Metal Compounds with Carboxylic Acid Group(s). J. Am. Chem. Soc. 2016, 138, 3891−3903. (30) Casarin, L.; Swords, W. B.; Caramori, S.; Bignozzi, C. A.; Meyer, G. J. Rapid Static Sensitizer Regeneration Enabled by Ion Pairing. Inorg. Chem. 2017, 56, 7324−7327. (31) Braslavsky, S. E. Glossary of terms used in photochemistry, 3rd edition (IUPAC Recommendations 2006). Pure Appl. Chem. 2007, 79, 293−465. (32) Yonemoto, E. H.; Riley, R. L.; Kim, Y. I.; Atherton, S. J.; Schmehl, R. H.; Mallouk, T. E. Photoinduced electron transfer in covalently linked ruthenium tris(bipyridyl)-viologen molecules: observation of back electron transfer in the Marcus inverted region. J. Am. Chem. Soc. 1992, 114, 8081−8087. (33) Maurer, A. B.; Hu, K.; Meyer, G. J. Light Excitation of a Bismuth Iodide Complex Initiates I−I Bond Formation Reactions of Relevance to Solar Energy Conversion. J. Am. Chem. Soc. 2017, 139, 8066−8069. (34) Li, G.; Brady, M. D.; Meyer, G. J. Visible Light Driven Bromide Oxidation and Ligand Substitution Photochemistry of a Ru Diimine Complex. J. Am. Chem. Soc. 2018, 140, 5447−5456. (35) Li, G.; Swords, W. B.; Meyer, G. J. Bromide Photo-oxidation Sensitized to Visible Light in Consecutive Ion Pairs. J. Am. Chem. Soc. 2017, 139, 14983−14991. (36) Li, G.; Ward, W. M.; Meyer, G. J. Visible Light Driven Nanosecond Bromide Oxidation by a Ru Complex with Subsequent Br−Br Bond Formation. J. Am. Chem. Soc. 2015, 137, 8321−8323. (37) Wehlin, S. A. M.; Troian-Gautier, L.; Li, G.; Meyer, G. J. Chloride Oxidation by Ruthenium Excited-States in Solution. J. Am. Chem. Soc. 2017, 139, 12903−12906.
This research was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award DE-SC0013461. L.T.-G. acknowledges the Belgian American Educational Foundation (BAEF) and the Bourse d’Excellence Wallonie-Bruxelles (WBI.World) for generous support. W.B.S. acknowledges support from the National Science Foundation Graduate Research Fellowship Program under Grant DGE-1650116.
(1) Yang, W. S.; Park, B.-W.; Jung, E. H.; Jeon, N. J.; Kim, Y. C.; Lee, D. U.; Shin, S. S.; Seo, J.; Kim, E. K.; Noh, J. H.; Seok, S. I. Iodide management in formamidinium-lead-halide-based perovskite layers for efficient solar cells. Science 2017, 356, 1376−1379. (2) Ardo, S.; Meyer, G. J. Photodriven heterogeneous charge transfer with transition-metal compounds anchored to TiO2 semiconductor surfaces. Chem. Soc. Rev. 2009, 38, 115−164. (3) Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. DyeSensitized Solar Cells. Chem. Rev. 2010, 110, 6595−6663. (4) Carey, G. H.; Abdelhady, A. L.; Ning, Z.; Thon, S. M.; Bakr, O. M.; Sargent, E. H. Colloidal Quantum Dot Solar Cells. Chem. Rev. 2015, 115, 12732−12763. (5) Ardo, S.; Park, S. H.; Warren, E. L.; Lewis, N. S. Unassisted solar-driven photoelectrosynthetic HI splitting using membraneembedded Si microwire arrays. Energy Environ. Sci. 2015, 8, 1484− 1492. (6) Kosower, E. M. The Effect of Solvent on Spectra. I. A New Empirical Measure of Solvent Polarity: Z-Values. J. Am. Chem. Soc. 1958, 80, 3253−3260. (7) Svensson, P. H.; Kloo, L. Synthesis, Structure, and Bonding in Polyiodide and Metal Iodide−Iodine Systems. Chem. Rev. 2003, 103, 1649−1684. (8) de Violet, P. F. Polyhalide radical anions as intermediates in chemistry. Rev. Chem. Intermed. 1981, 4, 121−169. (9) Demas, J. N.; Addington, J. W. Luminescence quenching of the tris(2,2’-bipyridine)ruthenium(II) and tris(1,10-phenanthroline)ruthenium(II) cations. J. Am. Chem. Soc. 1976, 98, 5800−5806. (10) Clark, C. C.; Marton, A.; Meyer, G. J. Evidence for Static Quenching of MLCT Excited States by Iodide. Inorg. Chem. 2005, 44, 3383−3385. (11) Farnum, B. H.; Gardner, J. M.; Marton, A.; Narducci-Sarjeant, A. A.; Meyer, G. J. Influence of ion pairing on the oxidation of iodide by MLCT excited states. Dalton Trans. 2011, 40, 3830−3838. (12) Farnum, B. H.; Jou, J. J.; Meyer, G. J. Visible light generation of I−I bonds by Ru-tris(diimine) excited states. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 15628−15633. (13) Gardner, J. M.; Abrahamsson, M.; Farnum, B. H.; Meyer, G. J. Visible Light Generation of Iodine Atoms and I−I Bonds: Sensitized I− Oxidation and I3− Photodissociation. J. Am. Chem. Soc. 2009, 131, 16206−16214. (14) Gardner, J. M.; Giaimuccio, J. M.; Meyer, G. J. Evidence for Iodine Atoms as Intermediates in the Dye Sensitized Formation of I− I Bonds. J. Am. Chem. Soc. 2008, 130, 17252−17253. (15) Marton, A.; Clark, C. C.; Srinivasan, R.; Freundlich, R. E.; Narducci Sarjeant, A. A.; Meyer, G. J. Static and Dynamic Quenching of Ru(II) Polypyridyl Excited States by Iodide. Inorg. Chem. 2006, 45, 362−369. (16) Rowley, J. G.; Farnum, B. H.; Ardo, S.; Meyer, G. J. Iodide Chemistry in Dye-Sensitized Solar Cells: Making and Breaking I−I Bonds for Solar Energy Conversion. J. Phys. Chem. Lett. 2010, 1, 3132−3140. (17) Swords, W. B.; Li, G.; Meyer, G. J. Iodide Ion Pairing with Highly Charged Ruthenium Polypyridyl Cations in CH3CN. Inorg. Chem. 2015, 54, 4512−4519. 178
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
Article
Accounts of Chemical Research (38) Brady, M. D.; Troian-Gautier, L.; Sampaio, R. N.; Motley, T. C.; Meyer, G. J. Optimization of Photocatalyst Excited- and GroundState Reduction Potentials for Dye-Sensitized HBr Splitting. ACS Appl. Mater. Interfaces 2018, 10, 31312−31323. (39) Sheridan, M. V.; Wang, Y.; Wang, D.; Troian-Gautier, L.; Dares, C. J.; Sherman, B. D.; Meyer, T. J. Light-Driven Water Splitting Mediated by Photogenerated Bromine. Angew. Chem. 2018, 130, 3507−3511. (40) Barigelletti, F.; Juris, A.; Balzani, V.; Belser, P.; Von Zelewsky, A. Excited-state properties of complexes of the tris(diimine)ruthenium(2+) ion family. Inorg. Chem. 1983, 22, 3335−3339. (41) Meyer, T. J. Photochemistry of metal coordination complexes: metal to ligand charge transfer excited states. Pure Appl. Chem. 1986, 58, 1193−1206. (42) Motley, T. C.; Troian-Gautier, L.; Brennaman, M. K.; Meyer, G. J. Excited-State Decay Pathways of Tris(bidentate) Cyclometalated Ruthenium(II) Compounds. Inorg. Chem. 2017, 56, 13579−13592. (43) O’Donnell, R. M.; Johansson, P. G.; Abrahamsson, M.; Meyer, G. J. Excited-State Relaxation of Ruthenium Polypyridyl Compounds Relevant to Dye-Sensitized Solar Cells. Inorg. Chem. 2013, 52, 6839− 6848. (44) Troian-Gautier, L.; Moucheron, C. RutheniumII Complexes bearing Fused Polycyclic Ligands: From Fundamental Aspects to Potential Applications. Molecules 2014, 19, 5028. (45) Moucheron, C.; Kirsch-De Mesmaeker, A.; Kelly, J. M. Photoreactions of ruthenium(II) and osmium(II) complexes with deoxyribonucleic acid (DNA). J. Photochem. Photobiol., B 1997, 40, 91−106.
179
DOI: 10.1021/acs.accounts.8b00373 Acc. Chem. Res. 2019, 52, 170−179
